Bachelor Thesis Open Access
Bertsch, Johannes
<?xml version='1.0' encoding='utf-8'?> <oai_dc:dc xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/oai_dc/ http://www.openarchives.org/OAI/2.0/oai_dc.xsd"> <dc:creator>Bertsch, Johannes</dc:creator> <dc:date>2025-01-31</dc:date> <dc:description>In this work, different approaches to utilizing Transformer-based models for classification are tested, with efforts made to improve the performance of current methods, such as FastBDT. Different models (TabTransformer, FT-Transformer, TabNet and Event Classifier Transformer) are tested, and one model, the Event Classifier Transformer, is selected for a detailed performance study. The classification for the B+→K∗+τ+τ− analysis by Lennard Damer, which employs the FastBDT model, is redone using the Event Classifier Transformer on the same dataset, and the results are compared. To allow for a fair comparison, all steps involved in training and hyperparameter optimization are replicated from the original analysis. The sensitivity of the analysis is evaluated by estimating the upper limit of the branching fraction of the process of interest. Additionally, an algorithm for finding the feature importance is employed.</dc:description> <dc:identifier>https://publish.etp.kit.edu/record/22290</dc:identifier> <dc:identifier>oai:publish.etp.kit.edu:22290</dc:identifier> <dc:language>eng</dc:language> <dc:relation>url:https://publish.etp.kit.edu/communities/belle2</dc:relation> <dc:relation>url:https://publish.etp.kit.edu/communities/etp</dc:relation> <dc:rights>info:eu-repo/semantics/openAccess</dc:rights> <dc:subject>Belle II</dc:subject> <dc:subject>Transformer</dc:subject> <dc:title>Event Selection for $\mathrm{B}^+ \rightarrow \mathrm{K}^{*+} \tau^+ \tau^-$ Using Transformers at Belle II</dc:title> <dc:type>info:eu-repo/semantics/bachelorThesis</dc:type> <dc:type>thesis-bachelor-thesis</dc:type> </oai_dc:dc>